• VALIDATION
  • DEPLOYMENT
  • HUMAN
  • LOGS

How to deploy new indexing pipelines with your AI

Connect Goldsky to your agent to automate Web3 data indexing. Your AI handles the configuration checks and the deployment itself.

Ask AI about this page

Short answer

How can I automate Web3 data indexing with my AI?

Your agent uses Goldsky tools to check your JSON configs for errors before pushing them live. This ensures your subgraphs and data pipelines are valid before they hit the network. You get a reliable deployment loop without manual config debugging.

Deployment outcomes

Where your data lands.

What happens after you give your agent a configuration.

VALIDATION

Error-free configs

The AI checks your JSON against Goldsky requirements. You avoid deployment failures caused by syntax errors or missing fields.

DEPLOYMENT

Live pipelines

Once the config is clean, your agent pushes the pipeline. Your data indexing starts running on the network immediately.

HUMAN

Manual oversight

You decide when to trigger a push. The AI won't deploy anything without your direct command.

LOGS

Deployment status

The agent reports back on whether the pipeline was accepted or rejected. You get clear feedback on why a config failed.

The workflow

What your AI does when a new config is ready.

The agent moves from checking your work to executing the deployment.

  1. Check JSON integrity

    The agent inspects your indexing configuration to ensure it meets Goldsky's schema requirements.

    validate_pipeline

  2. Fix configuration errors

    If the validation fails, the agent identifies the specific field causing the issue. You can then correct the JSON based on its feedback.

    validate_pipeline

  3. Push to Goldsky

    After the check passes, the agent sends the validated configuration to the Goldsky platform.

    create_pipeline

  4. Confirm deployment

    The agent verifies the pipeline was successfully created and notifies you that the indexing is active.

    create_pipeline

Try it

Copy these to start.

Use these prompts to kick off your first deployment.

Starting points

Replace the bracketed text with your actual JSON or specific pipeline names.

Goldsky Connector

You're all set. Choose your MCP client and follow the setup instructions.

Connector linkhttps://edge.vinkius.com/vk_preview_vKwCn24rK9jiYRm7bJaactLmNUpPZSSNAnWrQ0l6/mcp

Claude Desktop

Follow the steps below to connect in seconds.

  1. 1In Claude Desktop, open Settings → Connectors.
  2. 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
  3. 3Click Add and start a new chat — Goldsky capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "goldsky-web3-data-indexing-subgraphs-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_vKwCn24rK9jiYRm7bJaactLmNUpPZSSNAnWrQ0l6/mcp"
    }
  }
}
Copy into chat04

  • Validate this JSON configuration for a new Goldsky pipeline: [paste JSON here]

  • Create a new indexing pipeline using this configuration: [paste JSON here]

  • Check if my current subgraph configuration is valid for deployment.

  • Deploy a new data pipeline with the following parameters: [list parameters]

  • Claude
  • ChatGPT
  • Cursor
  • VS Code
  • Windsurf
  • Claude Code
  • JetBrains
  • Cline

Start here

Connect Goldsky once, then ask.

Link your Goldsky account to Vinkius. Your credentials stay encrypted, and you can immediately start managing pipelines through your AI client.

Connect Goldsky to your AI

FAQ

How the deployment process behaves

  • 01

    Can my AI deploy a pipeline without me seeing it first?

    No. The agent only performs the deployment when you explicitly tell it to push the configuration.

  • 02

    Will the AI fix my JSON errors automatically?

    The AI can identify errors and suggest fixes, but it won't rewrite your files or push changes without your permission.

  • 03

    Does the AI handle the actual data indexing?

    No. The AI manages the deployment and configuration. Goldsky handles the actual data processing and indexing once the pipeline is live.

  • 04

    Can the AI delete existing pipelines?

    No. The current tools only allow for validating and creating new pipelines.

  • 05

    What happens if the validation fails?

    The agent will stop the process and tell you exactly what is wrong with the JSON so you can fix it.

  • More questions about Goldsky? The Connector page answers them. See everything the Goldsky Connector can do

Connect Goldsky to Claude, Cursor, ChatGPT & more